Of the 243 patients who received a PKB between October 1, 2018, and October 1, 2023, at our institute, 30 were excluded based on the exclusion criteria, and 213 were ultimately analyzed. Patient characteristics are shown in Table 1. Of the 213 patients, 110 (51.6%) were men, 28 (13.1%) had diabetes, and 102 (47.4%) had hypertension. The median (IQR) values for the age, body mass index (BMI), and eGFR were 56 (40–71) years old, 22.3 (20.2–25.0) kg/m2, and 53.6 (33.7–72.8) mL/min/1.73m2, respectively. The median (IQR) blood pressure at the beginning of the biopsy was 136/80 (122–153/70–90) mmHg. The needles were used for a median (IQR) of 3 (3–4) passes.

The histopathological diagnoses of the PKBs are shown in Table 2. The most frequent histopathological diagnosis was immunoglobulin A (IgA) nephropathy (n = 72, 34.0%), followed by membranous nephropathy (n = 17, 8.0%) and tubulointerstitial nephritis (n = 14, 6.6%). Biopsy specimens from 4 patients were inadequate for a histopathological diagnosis due to too few glomeruli (< 7 glomeruli [20]).

Fifty-four (25.3%) patients experienced Hb loss ≥ 1 g/dL due to a PKB. Only 2 (0.9%) patients received blood transfusions after the biopsy, with an Hb loss of 2.2 and 3.4 g/dL reported. The patient with 3.4 g/dL Hb loss underwent a computed tomography (CT) scan, which showed a large hematoma (estimated volume of 155 mL). Other patients were found to have no or mild hematoma on routine ultrasound the day after PKB. Thirteen patients (6.1%) had macroscopic hematuria after PKB, which improved with supportive care, which was continued until the next morning. No post-biopsy interventional radiology or nephrectomy was required to control hemorrhaging. None of the patients died due to a PKB in this case-control study.

The incidence rate of Hb loss ≥ 1 g/dL did not markedly differ among the stages of chronic kidney disease (p = 0.81, Fig. 2a). Patients who were clinically diagnosed with nephrotic syndrome and acute kidney injury (AKI) before the biopsy seemed to have a high incidence rate of Hb loss ≥ 1 g/dL, although there were no statistically significant differences among clinical diagnoses before the biopsy (p = 0.09, Fig. 2b).

Incidence rate of hemoglobin (Hb) loss ≥ 1 g/dL after percutaneous kidney biopsy. The comparison of the incidence rate of Hb loss ≥ 1 g/dL among chronic kidney disease (CKD) stages (a) and clinical diagnoses before a biopsy (b). There were no differences in rates among CKD stages or clinical diagnoses. The numbers on the top of the bar graph represent the incidence rate as a percentage. Data were analyzed using Fisher’s exact test. AKI: acute kidney injury, CKD: chronic kidney disease

Next, we investigated the predictive ability of the urinary NAG/Cr ratio for the incidence of Hb loss ≥ 1 g/dL due to a PKB using an ROC curve. The ROC curve showed that the urinary NAG/Cr ratio could predict the incidence of Hb loss ≥ 1 g/dL, with an AUC of 0.65 (p = 0.005, Fig. 3a). In contrast, the urinary β2MG level did not predict the incidence of Hb loss ≥ 1 g/dL (AUC 0.46, p = 0.49; Fig. 3b). The optimal cutoff value of the urinary NAG/Cr ratio for predicting the incidence of Hb loss ≥ 1 g/dL was 35 U/L.

Receiver operating characteristics (ROC) curves. (a) Urinary N-acetyl-β-D-glucosaminidase (NAG)/Cr, (b) urinary β2-microgrobulin (β2MG). AUC: area under the receiver operating characteristic curve

We further examined whether a high urinary NAG/Cr ratio was a risk factor for the incidence of Hb loss ≥ 1 g/dL. In this study, we defined a high urinary NAG/Cr ratio as ≥ 35 U/L. The incidence of Hb loss ≥ 1 g/dL was higher in patients with dyslipidemia (odds ratio [OR] 1.95, 95% confidence interval [CI] 1.02–3.77, p = 0.047), abnormal APTT (OR 3.93, 95% CI 1.71–9.03, p = 0.001), urinary protein ≥ 3 g/gCr (OR 2.61, 95% CI 1.38–4.94, p = 0.003), and a high urinary NAG/Cr (OR 4.69, 95% CI 2.27–9.70, p < 0.0001; Table 3) than in others.

A multivariable logistic regression analysis showed that abnormal APTT (OR 3.15, 95% CI 1.42–7.27, p = 0.005) and a high urinary NAG/Cr ratio (OR 3.21, 95% CI 1.42–7.27, p = 0.005) were independent risk factors for the incidence of Hb loss ≥ 1 g/dL (shown in Table 3).

In addition, another multivariable logistic regression model that used previously reported risk factors also indicated that a high urinary NAG/Cr was an independent risk factor for the incidence of Hb loss ≥ 1 g/dL (OR 5.32, 95%CI 2.46–11.5, p < 0.0001) (Table 4).

To explore why a high urinary NAG/Cr ratio was a risk factor for Hb loss ≥ 1 g/dL, we additionally evaluated the pathological findings of PKBs. The associations between the pathological findings and urinary NAG/Cr values are shown in Fig. 4. The urinary NAG/Cr values were significantly higher in patients with moderate-to-severe multilayered elastic laminae of the small muscular artery (p = 0.008, Fig. 4a), moderate-to-severe tubular injury (p = 0.005, Fig. 4b), moderate-to-severe fibrosis (p = 0.004, Fig. 4c), and infiltration of inflammatory cells (p = 0.04, Fig. 4d) than in those who had none or only mild pathological disorders.

Comparisons of urinary NAG/Cr ratios by the severity of pathological findings. (a) Multilayered elastic lamina of small muscular artery, (B) tubular injury, (c) fibrosis, and (d) inflammatory cell infiltration. Patients were divided into two severity groups (none/mild vs. moderate/severe) for each pathological finding. Dots represent each individual. Box-and-whisker plots: within each box, horizontal lines denote the median values; boxes extend from the 25th to the 75th percentile of each group’s distribution of values; vertical extending lines denote adjacent values (i.e. the most extreme values within the 1.5 interquartile range [IQR] of the 25th and 75th percentile of each group). Data were analyzed using the Mann-Whitney U-test. NAG, N-acetyl-β-D-glucosaminidase

The associations between pathological findings and the incidence rates of Hb loss ≥ 1 g/dL are shown in Fig. 5. The incidence rates of Hb loss ≥ 1 g/dL were significantly higher in patients who had moderate-to-severe multilayered elastic laminae of the small muscular artery than in those who had none or mild one (28.5% vs. 15.2%, p = 0.03, Fig. 5a). However, the severity of other pathological findings, such as fibrosis, tubular injury, and inflammatory cell infiltration, were not associated with an elevated bleeding risk.

Comparisons of the incidence rate of hemoglobin (Hb) loss ≥ 1 g/dL by severity of pathological findings. (a) Multilayered elastic lamina of small muscular artery, (b) tubular injury, (c) fibrosis, and (d) inflammatory cell infiltration. Patients were divided into two severity groups (none/mild vs. moderate/severe) for each pathological finding. The numbers on the top of the bar graph represent the incidence rate as a percentage. Data were analyzed using Fisher’s exact test